Overview
Prompt-injection check for untrusted text — Scan untrusted text (web page, email, tool description, agent message) for prompt-injection shapes — instruction override, role redefinition, credential-moving requests, hidden tags, invisible Unicode. {text} up to 8,000 chars. Verdict + findings; pattern-based. (MCP tool: prompt_injection_check).
Health
x402 Payment Validation
Recent Health Checks
| Time | Status | HTTP | Latency | Error |
|---|---|---|---|---|
| 2026-10-07 04:27:56 | healthy | 402 | 69ms | |
| 2026-10-07 00:38:05 | healthy | 402 | 60ms | |
| 2026-10-06 12:24:22 | healthy | 402 | 51ms | |
| 2026-10-05 22:55:20 | healthy | 402 | 61ms |